 Live from Vancouver, Canada, it's theCUBE. Covering OpenStack Summit North America 2018. Brought to you by Red Hat, the OpenStack Foundation, and its ecosystem partners. Welcome back to theCUBE's coverage of OpenStack Summit 2018 in Vancouver. I'm Stu Miniman with my co-host for the week, John Troyer. Happy to welcome to the program first time guests, Ildiko, Vansha, coming off the Edge keynote presentation this morning. She is the ecosystem technical lead with the Edge Computing Group as part of the OpenStack Foundation. Thanks so much for joining us. Thank you. Yeah, it's coming into the show, Edge is one of those things that it was actually pretty exciting to talk about because Edge is not only super hot, but when I thought back to previous shows, this is the sixth year we've had theCUBE here, and my fifth year doing it, it's like, wait, I've been talking all the telcos for years here, NFV was one of those use cases, and when you connect the dots, it's like, oh, Edge, of course. You know, I said, you know, this conference is actually, you know, hipster when it comes to, you know, Edge, we were totally covering it, you know, well before we called it that. So, you know, explain to us, you know, your role in the foundation and what led to kind of the formation of this track. Yes, so I'm the ecosystem technical lead within the foundation, which is basically a role that belongs under the business development team, so I'm basically building connections with our ecosystem members. I'm trying to help them succeed with OpenStack, both as software package and as a community. We are embracing open source, of course, so I'm also trying to advocate for involvement in open source because I think that's a key, like, you know, picking up an open source software component and use it, that's a great start, but if you really want to be successful with it and you want to be able to successfully build it into your business model then getting involved in the community, you know, both enhancing the software and maintaining the software, that's really key. So my role is also like onboarding companies as well to be active members of the community and my focus is shifting towards edge computing. The history of edge computing in OpenStack basically started last May when Beth Cohen from Verizon described their use case, which is, you know, OpenStack in a tiny box in production, it's like, oh, wow. So that was also a little bit of an eye-opener for us as well that yes, it's telecom, it's 5G, but this is the thing that's called edge and maybe this is something that we should also, you know, look deeper into. So we went to San Francisco last September, OpenDev, 200 people, architects, software developers, trying to figure out what edge computing is we had. I think we had the question of every single session, someone asked that, okay, yeah, so what did you mean exactly when you said edge? Because, you know, from the nature of the architecture, like you have the central cloud and then the sites on the different barriers or levels. There are several edges depending on how far you want to go. So, for you and OpenStack, what does edge mean or all the above? With OpenStack, so after OpenDev, when we realized that it's not really a well-defined term, we wrote up a white paper, it's at OpenStack.org slash edge. It's a short one really to just set the ground for what edge computing is and what we came up with is, so don't imagine like a two-sentence definition for edge computing because I still strongly believe that doesn't exist and anyone who claims it, they, that's not true. So what we did with the white paper is basically, we set characteristics and criterias that defines cloud edge computing per se. Like, you know, what people are talking about when you're moving out the compute and never getting closer to the edge, like what that means from the bandwidth perspective, from how you will manage it, what that means for security and all these sort of things and you can basically characterize what edge means. So we rather described these layers and how far we go and as far as like the edge, the very end edge device and like the IoT sensors, that's not a target of OpenStack. So OpenStack itself is infrastructure as a service, so our edge computing group is still staying on that layer. The edge computing group itself is focusing on the angles what edge brings onto the table, all these requirements, collecting the use cases and trying to figure out what's missing, what we need to implement for the next steps. If I can repeat and maybe I'll get it right or wrong, right, the idea is at a cell tower or at a remote office or branch office or some closet somewhere, there's a full set of OpenStack running, maybe a minimal set of OpenStack, but it's live, it's updatable, you can update services on it, you can update the actual OpenStack itself and it doesn't need bespoke hardware necessarily, but it's now updatable and part of a bigger multi-cloud infrastructure from some sort of service entity or enterprise, is that fair? I think that's fair, I mean, so there's OpenStack itself that people know very well, a lot of projects. So when we talk about edge, obviously we don't want to say that okay, pick the whole thing and install all the 60 projects because that's really not suitable for edge. So for example, the group is looking into that, which OpenStack components are essential for edge and also the group is defining small edge, medium edge, what that means from hardware, footprint perspective, so just to figure out what the opportunities are there, what will fit, what will not fit, so OpenStack itself is very modular by today, so you can pick up the services that you need. So what we discussed for example this week is keystone, identity, you need it, of course. So how much that fits into the edge scenarios and I think the main conclusion of the forum session yesterday was that yeah, keystone supports federation, we talked through the cases and it seems like that it's kind of there, so we now need a few people who will sit down, put together the environment and start testing it because that's when it comes out that almost there but there are a few things to tweak but basically the idea is what you described, pick up the component and put it there and work with it. We also have like another project called Cyberg which is fairly new, that's for hardware acceleration, so it is providing a framework to plug in GPUs, FPGAs and these sort of a bit more specialized hardware which will be really useful for edge use cases to OpenStack, so that's for example something that China Mobile and the OPNFV edge cloud group is looking into to use so I really hope that we will get there this year to test it in the OPNFV files labs in action, so we also have pretty great cross-community collaboration on trying to figure this whole thing out. It often helps if we have examples to talk about, to really explain this, Beth Cohen we spoke with her last year and absolutely caught our attention, got a lot of feedback from the community on it, had Contron on earlier this week talking about, John was saying, here's some small device there with a little blade and running pieces of OpenStack there to be able to run, anything from the keynote or you know, I think there's 40 sessions that you've got here, if you can give us a couple of kind of examples of some of the use cases that we're seeing to kind of bring this edge to reality. Example use cases is we just heard this morning, for example, someone from the textile industry like how to detect issues with the fabric, so this is like one new manufacturing use case. I also heard another one which is not checking the fabric itself, but basically the company who manufactures those machines that they are using to create the fabric, so they would like to, you know, have a central cloud and have it connected to the factories, so being able to monitor, you know, how the machines are doing, how they can improve those machines, and also within the factory to monitor all the circumstances, because for all the chemical processes, it's really important that the temperature and everything else is just, you know, clicks because otherwise all your fabrics will have to go to trash, so that's manufacturing. A lot of telecom 5G, obviously that is really, really heavy because that's the part of the industry which is there today, so with 5G, all those strict requirements, this is really what we are mainly focusing on today. We are not specializing anything for telecom and 5G use cases, but we want to make sure that all our components fit into that environment as well. In the white paper, for example, you also could see the retail use case. I'm not sure whether that will be exactly on stage this week, but that is also a great example on like Walmart with a lot of stores around, so how you manage those stores, because they're also not wanting to do everything centrally, so they would like to move the functionality out, what if the network connectivity is cut, they still have to be able to operate the store as nothing happened, so there are a lot of segments of the industry who already have kind of really well-defined use cases and what we see is that there's many overlapping between the requirements from the different segments that we can purchase address. Are we seeing things like AI and ML coming up in these conversations also? Yes, I think it was the manufacturing use case when I heard that they are planning to use that and it's popping up. I think as far as our group is concerned, we are more looking into, I don't know, let's say lower-level requirements, like how you maintain and operate the hundreds and thousands of S sites, what happens with security, what happens with monitoring, what happens with all these sort of things, like we have a new project rolling in under the foundation umbrella called Airship, which is basically deployment and lifecycle management, which supposed to address one of the aspects that you were talking about on, okay, so how you manage this, how you upgrade this. And upgrade is, again, a really interesting question because I think I talked to someone yesterday who was like, yes, the Contra guys, they were saying that, yeah, upgrade, it's really ambitious, so let's say that maybe 18, 24 months or something like, when the tech operator will decide to upgrade something out in the edge because it's out there, it's working, let's not touch this. So when we talk about upgrade, even that I think will depend on the bits of the industry that what pace they will decide to take in. Are there any particular surprises or learnings that you've had this year after talking with this community for a week now? You said, well, last year, I was very impressed last year when they got up on stage and talked about that, that kind of expanded my mind a little bit. You've been working with this now for a year, there's a whole track and forum sessions. Anything you're excited about taking to the future or learnings or surprises that, oh, this is really going to work or anything like that, any parts of it that are really interesting? You talked about security upgrades, I mean, we've talked about a lot of the technical components, but it seems like it's working. I think at this point, at least on my end, I think I'm over the surprise phase, so what surprises me the most is how many groups that are out there who are trying to figure out what this whole edge thing is and what we need to really focus on among the technical requirements is that how we are working together with all these groups just to make sure that the integration between the different things that we are all developing and working on is smooth. So like we've been working together with the OP and FE community for a while now, it's a really fruitful relationship between us, like seeing OpenStack being deployed in a full-stack environment and being tested, that's really priceless, and we are planning to do the same thing with Edge as well, and we are also looking into Onab, Aquino, Etsy Mac, so looking into the open-source groups, looking into the standardization and really just trying to ensure that when we talk about open infrastructure that really is designed and developed in a way that integrates well with the other components, it's synchronized with the standardization activities because I think especially in case of Edge, when we say interoperability, that's a level higher than what we called interoperability on the telecom level, I think, like when you just imagine one operator network and applications from other providers popping up in that network and components that just realizing the network popping up from different vendors and this whole thing has to work together, so I think OpenStack and Open Infrastructure has a really big advantage there compared to any proprietary solution because we have to address this, I think, really big challenge and it's also a really important challenge. Elika, really appreciate giving us all the updates here on the Edge track, the keynote, definitely one of the areas that capturing our attention in lots of people out there, so thanks so much for joining us. Thank you for the opportunity. All right, for John Troyer, I'm Stu Miniman, lots more coverage here from the OpenStack Summit 2018 in Vancouver, thanks for watching theCUBE.